Jantjaet 22, 1909] 



SCIENCE 



125 



buildings, and shown its close correlation 

 with the distribution of underlying ma- 

 terial;* and similar, though less detailed, 

 correlations have been made in other 

 regions. On the theoretic side the subject 

 is almost untouched, and there is great 

 need of experimentation, but the empiric 

 results already available have much practi-; 

 cal value and enable the geologist and engi- 

 neer to distinguish broadly, within the 

 limits of a malloseismie district, the tracts 

 more likely, and the tracts less likely, to be 

 affected disastrously by the passing earth- 

 quake wave. 



On the whole that factor of earthquake 

 forecast which consists in the indication of 

 locality is in a satisfactory condition. In 

 long inhabited regions experience desig- 

 nates certain districts as malloseismie. 

 Newly settled regions may be classified, 

 provisionally and less perfectly, by the 

 data of physiography. And malloseismie 

 districts will eventually be subdivided 

 with confidence by means of geologic 

 criteria. 



TIME 



Turning now to the time factor in fore- 

 casting, and retaining the point of view 

 which emphasizes the element of danger, 

 let us inquire what methods are available 

 for the prediction of the time of occur- 

 rence of a destructive earthquake at a 

 given locality or in a given district. Ra- 

 tional attempts to solve this problem have 

 been connected (1) with the idea of 

 rhythm, (2) with that of alternation, (3) 

 with that of the trigger or starter, and (4) 

 with that of the prelude; and each of 

 these lines of approach is worthy of ex- 

 amination. 



Rhythm. — Because we are surrounded 

 by and immersed in the rhythms of art 

 and nature, and because the earthquake is 



* " Report of the California Earthquake Com- 

 mission," Vol. I., Part II., pp. 220-45, and atlas, 

 maps 17-19. 



a recurrent phenomenon, it is easy to infer 

 that the interval between the last shock 

 and the next will be similar to that between 

 the last and its predecessor. Reasoning of 

 this general tenor probably underlies the 

 greater number of lay forecasts, and is in 

 particular responsible for the wide-spread 

 popular belief that a place recently de- 

 vastated is ipso facto immune for several 

 decades, or at least for several years. A 

 similar belief prevalent among men of 

 science has a slightly different origin, but 

 is even more strongly held; and there is 

 little exaggeration in saying that our guild 

 recognize it is a duty, when the terror- 

 stricken inhabitants of a racked and ruined 

 city seek safety in the open spaces, to as- 

 sure them that the danger is past and 

 urge them to return to their homes. Now, 

 it is not at all true that either the great 

 shocks or the small shocks affecting a par- 

 ticular locality, or affecting a district, or 

 affecting the earth as a whole, are sepa- 

 rated one from another by regular or ap- 

 proximately regular intervals; and it is 

 not at all true that immediate danger is 

 past when a great shock has wrought its 

 havoc ; and yet I am prone to believe that 

 the rhythmic principle does hold place in 

 the mechanics of earthquakes. On that 

 point something further will be said, but I 

 shall first invite your attention to the gen- 

 eral phenomena of earthquake sequence, 

 selecting examples from the American 

 record because we are most interested in 

 the phenomena of our ovm territory. 



The United States has one well-known 

 malloseismie district, a district including 

 central and southern California, with 

 areas in Mexico and the Pacific Ocean, and 

 possibly extending northward. Alaska 

 also contains a district, and there may be 

 a third in Utah. Since the beginning of 

 the last century, Alaska has experienced 

 at least nine shocks of destructive rank; 

 but the record is fragmentary and may 



